Video games that simulate the experiences of combat, space travel and car theft have achieved a startling level of fluidity and detail in recent years to create increasingly realistic virtual worlds. When it comes to medicine, however, the graphics that doctors and surgeons have to work with are closer to the grainy, cartoonish images of the Atari generation than they are to the video games Assassin's Creed or Grand Theft Auto.

The computing power required to render virtually realistic organs and soft tissue is still unavailable to most physicians (except for a handful with access to supercomputers), but it's coming, says Joseph Teran, an assistant mathematics professor at the University of California, Los Angeles.

Teran envisions a time within the next five years when medical professionals will be able to scan patients prior to procedures and create three-dimensional virtual images of their bodies, which they can store in computers and use for practice before performing the real surgeries.

Tissue, muscle and skin are elastic and behave like a spring, and their characteristics can be expressed using classical mathematical theory, Teran says.

To develop virtual models of patients, physicians must create geometric representations of their tissue and organs using either magnetic resonance imaging (MRI) or computed tomography (CT). The information that these scanners currently provide, however, is in the form of numbers representing shades of gray, which are insufficient for creating accurate, real-color, three-dimensional renderings. "It's a data processing problem, because this information is not in the correct format," Teran says. He also estimates that it would now take as many as 20 professionals up to nine months to produce even a roughly accurate model of the human body.

The answer? Faster computers with improved software algorithms that can solve mathematical problems that have unknown elements and multiple independent variables. Teran is confident that the technology will catch up with the demands placed on it, and that researchers must be ready for the day when this improved technology becomes available. "Even if you don't have eight-core processors now, you can use large computing clusters at universities to test your algorithms," he says. Some of his colleagues are even working with chipmaker Intel to make sure their software will work with future generations of Intel technology.

To date, virtual surgery models have been primarily used to create before and after images in reconstructive surgery procedures. Other uses have been limited to simulating specific body parts, such as an organ, a cluster of muscles or a craniofacial malformation. In reconstructive surgery, virtualization has been used to map movement disorders like those associated with cerebral palsy, allowing surgeons to, via a computer, rearrange a patient's muscular skeletal model to de-emphasize weaker muscles and emphasize stronger ones before setting foot in an actual operating room. Virtual surgery used in this manner has also been limited by technology, as most computers available to surgeons for this work can do little more than render relatively crude images of tissue, muscles and organs. "It's something that's there but more at a limited scale," Teran says.

1 Comments

1. 1. nwjohn 10:10 AM 12/8/07

   The time predictions in this article are too pessimistic. Much of the technology required is here now. In the UK we already have prototype systems being developed by multi-disciplinary teams (clinicians, computer scientists, physicists, psychologists, etc). We will have a patient specific virtual environment trainer for interventional radiology deployed in a cost concious hospital in 2 years time, not 20.
   
   Prof. Nigel John
   Bangor University, UK

   Reply | Report Abuse | Link to this